Dual 4870x2 on PCI-E 2.0 8x performance loss?

Just wondering if there was a thread with someone that did a test to see just how much of a loss in performance there was when using these 2 cards in dual 8x 2.0 vs dual 16x 2.0? If there is a thread already dedicated to this or a site that shows this then please link.

Thanks

There is a loss with 4850 xfire and high res and/or af/aa turned up.

http://www.tweaktown.com/articles/1472/1/intel_p45_vs_x48_crossfire_performance/index.html

http://www.ocforums.com/showthread.php?t=571156
its in the sticky :O

Thats on the 4870, not the X2.

Yea I see there is very very little difference when xfiring 2 4870's on 8x 2.0 but I really wanted to see 2 4870x2's on 8x 2.0.

OBLIVIONLORD said:

Yea I see there is very very little difference when xfiring 2 4870's on 8x 2.0 but I really wanted to see 2 4870x2's on 8x 2.0.

Click to expand...

Brolloks does mention later on in the thread that he started to benchmark 4870x2 xfire and quit because he saw that they were bottlenecked and it wasn't worth it finishing.

It would have been nice to see someone benchmark COD4 when doing this. That game is very gpu intense. It only becomes CPU related when you get to quadfire/tri-sli, but quickly goes away at the higher res.

I'm playing it safe and going to run a X48 board for 4870x2 xfire. I think I'm getting bottlenecked with my pci-e 1.1 board and don't want to chance it with a P45 one. It would be nice to be able to run 16gb of memory though.

I did see his entry when he said "I did test it with dual 4870X2's btw and it crushed the x 8 lanes of the P45" as well as "I saw after the 1st few that the X2's were killing the x8 lanes and stopped" but i would really liked to have seen the results to actually see what the difference was. It stinks that he didnt post it.

I agree, if we were able to see the numbers, we could make a judgment call on getting x2 cards and P45 or not.....